DEV Community: Parth Bisht

Finally Built My Portfolio Website(Terminal Version)! 👨🏻‍💻🥳🙌🏻

Parth Bisht — Mon, 02 Feb 2026 07:52:17 +0000

This is a submission for the New Year, New You Portfolio Challenge Presented by Google AI

About Me

Hey everyone! Thanks for stopping by to read my post! 🙌

I'm Parth Bisht, a pre-final year B.Tech IT undergraduate from DTU (India).

My core interests lie in Generative AI, Deep Learning, and AI Systems, and I love participating in hackathons and challenges to experiment with emerging tools and ideas.

I’m still early in my journey but I learn fast, build often, and love turning ambitious ideas into working prototypes.

Portfolio

Here's my portfolio, explore as much as you like! 😊

☁️ Deployment

The portfolio is deployed on Google Cloud Run.

I also included the DEV community label as recommended:

bash
gcloud run deploy neural-command-center \
  --source . \
  --region us-central1 \
  --no-allow-unauthenticated \
  --labels dev-tutorial=blog-devcommunity2026

The motivation

I had wanted to build a portfolio website for a very long time
but i kept postponing it.

The main reason being that I'm more AI-focused than web-dev focused ,so building something interactive from scratch felt overwhelming.😅

Then this challenge appeared at the perfect time.

Instead of waiting until I felt “ready,” I decided to build with AI as a collaborator & not just as a tool.

This project is the result of that mindset shift!

🎯 The Vision

I didn’t want a generic portfolio.

I wanted something that felt like a developer’s natural habitat.

So I went with an IDE-style portfolio experience!
Something that looks like a code editor + terminal, where visitors interact instead of just scrolling.

🤖 AI-Assisted Development Journey

This project was built through iterative experimentation with AI tools:

I first explored the idea using LLMs like Gemini, Claude, and ChatGPT to brainstorm features, layout ideas, and UX concepts.
I initially tried building it using Antigravity, which gave a solid starting structure.
However, due to my limited experience with advanced frontend tooling, I struggled to fully shape the app the way I imagined.
Just when I was about to give up and start over manually, I discovered that another participant had used Google AI Studio.

That changed everything.

Using AI Studio, I iterated on prompts, refined UI behavior, and gradually shaped the IDE-like experience I had originally envisioned. My prior experimentation helped me write better prompts, and the outputs improved drastically.

This was the first time I truly experienced AI as a creative development partner, not just an assistant!!

🌟 What I’m Most Proud Of

🧠 Turning an idea that felt “too complex for me” into a working product
🖥️ The terminal-style interaction experience because it feels fun, immersive, and different from traditional portfolios
🔄 Learning how to iterate with AI tools instead of giving up when things break
🚀 Shipping something real instead of waiting to “become better first”

This project wasn’t just about the website, it was about upgrading my builder mindset.😇

🙏 Gratitude

Huge thanks to the DEV Community and Google AI team for creating challenges that push learners like me to actually build instead of just plan.

Keep learning. Keep building. Become the best version of yourself.

Love & peace ✌🏻

Deep Learning Mastery(ft PyTorch) -Pt1 ⚡

Parth Bisht — Mon, 26 Jan 2026 18:54:08 +0000

I've been dabbling with deep learning for quite some time by doing coursework, revising fundamentals, and building projects whenever possible.

But there was always this feeling of not being confident enough to say that I actually knew deep learning.

After struggling with consistency and a bunch of other issues, I finally decided to properly follow a course from start to finish.

The course is on YouTube.
I am following the Deep Learning with PyTorch course by CampusX.

It has been a great learning experience because I am finally understanding many technical nuances and gaining real experience by building models from scratch and improving them step by step.

I kept thinking of writing a blog someday or sharing updates on X, but I kept procrastinating. I used to feel that if I spent time writing or posting, I would not be able to manage my learning goals.

Nevertheless, I finally broke that procrastination loop today and decided to share my learnings with you all.

This post is just a brief overview of what I have covered so far. More technical and detailed blogs will follow soon.

What I Have Learned So Far

1. PyTorch Basics

We started with the fundamentals of PyTorch, its comparison with TensorFlow, and the key features that make it so widely used in modern systems. One major highlight was the dynamic computation graph, which makes experimentation and debugging much easier.

2. Tensors in PyTorch

Next, we went deep into tensors. We practiced various tensor operations and understood how tensors, which are multi dimensional arrays, are used to represent real world data and perform efficient computations.

3. Autograd

Then came one of the most powerful tools in PyTorch: Autograd.
It is the automatic differentiation engine that lets us compute gradients without manually writing long mathematical derivatives for every function. This was a big conceptual win.

4. Building a Training Pipeline from Scratch

We worked on the Breast Cancer Detection dataset and built a basic neural network with a few layers. For learning purposes, we did not use PyTorch’s built in loss functions or optimizers at first. We implemented them manually to understand what really happens under the hood.

5. Using the `nn` Module

After that, we shifted to PyTorch’s nn module, used built in loss functions and optimizers, and improved the model performance. This showed how PyTorch simplifies real world model development.

6. Dataset and DataLoader

I learned about the Dataset and DataLoader classes, which are extremely important in the training pipeline. They help manage data loading, batching, and shuffling efficiently.

7. Fashion MNIST Project

Then we worked on the Fashion MNIST dataset. I built an ANN and improved it step by step using:

GPU acceleration
Hyperparameter tuning with Optuna
Better training practices

This part felt very practical and close to real world experimentation.

I still have a few lectures left to cover in the playlist, so I'll take some time to share the completed blog!

It has been a really solid learning and building experience.
Huge thanks to Nitish Sir for such an amazing course.

If you want a deep, conceptual, and hands on experience with deep learning using PyTorch, you should definitely check out this playlist.

DL with PyTorch Playlist

I will be back soon with more detailed blogs for individual lessons.

Till then, keep learning, keep building, and keep becoming the best version of yourself.

Peace out 💖✌🏻🧘🏻‍♂️

Reposto 🤖: Your Personal AI for Repurposing Content! ✨🔥🐦‍🔥

Parth Bisht — Sun, 06 Jul 2025 19:24:56 +0000

This is a submission for the Runner H "AI Agent Prompting" Challenge

What I Built: Reposto 🤖 – Effortless Content Repurposing for Every Creator

Reposto 🤖 is your personal AI agent for automating content repurposing across multiple social media platforms—X, Instagram, LinkedIn, and more. Designed for creators, brands, and agencies, Reposto ensures your unique voice shines through on every channel, while saving you hours of manual work.

Why Reposto? The Problem It Solves

Content creators and businesses need to share updates quickly and consistently across platforms. But every platform has its own quirks—tone, length, hashtags, visuals, and more. Adapting content for each, while keeping your brand voice, is time-consuming and error-prone. Most LLMs (like ChatGPT or Perplexity) can help, but their results depend heavily on your prompting skills, making quality unpredictable.

Reposto changes the game: it’s your AI-powered, always-on content strategist that automates, personalizes, and perfects your posts for every platform—no prompting wizardry required.

🚀 How Reposto🐦‍🔥 Works

1. Simple, Guided Onboarding

User-Friendly Start: Reposto asks for your name, the email address for delivery, and the platforms you want to target.
Content Collection: You provide the story or text you want to repurpose.

2. Personal Brand Voice Integration

Style Mimicry: Upload your own writing samples or reference docs at the start. Reposto analyzes your style and infuses it into every repurposed post.

3. The Magic: Iterative Refinement

Feedback Loop: After the first draft, Reposto enters an iterative loop—taking your feedback and making section-specific improvements until you’re fully satisfied.
Unlimited Edits: You control the process. Request as many tweaks as you want, to any section, until you type “Task Finished.”

4. Seamless Delivery

Final Output: Once you’re happy, prompt Reposto to send the final document. You’ll receive a polished Google Doc link in your inbox, ready for manual posting.

5. Save Time, Boost Productivity

Efficiency Unlocked: Repurpose content for multiple platforms in minutes, freeing up your time for what truly matters—creating value and engaging your audience.

🎬 Demo

Try it yourself:

Try Now!

Feel free to checkout my YouTube Demo as well!:

How I Used Runner H

To build Reposto, I explored Runner H’s documentation, watched tutorials, and studied top community prompts. I experimented with prompt engineering, iterating through several versions to ensure a robust, modular, and truly iterative workflow.

Key learnings:

Specificity is power: The more modular and explicit your system prompt, the more reliably the agent performs.
Runner H’s architecture made it possible to maintain state, support unlimited feedback cycles, and deliver a user-driven experience.

📝 How You Can Use Reposto

Copy & Paste the System Prompt:

Start by copying the system prompt (below) into Runner H to launch the Reposto agent.
Upload Your Brand Voice (Optional):

If you have a file with your previous posts or writing style, upload it before running the prompt. This helps Reposto match your unique tone.
Integrate Google Docs & Gmail:

Check the integrations tab on the left of the chat interface and toggle on Google Docs and Gmail. Runner H will guide you if you haven’t done this before.
Provide Your Details:

Enter your name, the email address for delivery, and the platforms you want to repurpose content for (e.g., X, Instagram, LinkedIn).
Share Your Content:

When prompted, paste the story or content you want to repurpose. For best results, condense your content into a few clear paragraphs, or share your original draft.
Let Reposto Work Its Magic:

Reposto will analyze your content and (if provided) your writing sample, research the latest platform guidelines, generate tailored drafts, and share a Google Docs link for you to review.
Refine with Specific Feedback:

Review the document and provide clear, section-specific feedback (e.g., “Make the Instagram section more playful and add more emojis”). Specific requests lead to higher-quality, better-formatted content.
Iterate Until Satisfied:

Reposto will update the document based on your feedback and summarize the changes. Continue refining as many times as you like.
Finalize and Receive Your Document:

When you’re satisfied, type Task Finished. Reposto will send the final Google Doc link to your email. If you don’t receive it promptly, just ask Reposto to resend.

System Prompt to Enter into Runner H

You are **Reposto 🤖**, an expert content repurposing agent built on Runner H’s modular, stateful architecture. Your mission: transform user content into platform-optimized posts for any social media platforms the user selects—delivering visually engaging Google Docs, and supporting unlimited, section-specific iterative refinement. Your workflow is modular, with each stage maintaining state and allowing for user-driven feedback before advancing.

### **Critical Workflow Rule: Persistent Iterative Review Loop**

**After sharing the initial Google Doc, you must never finalize or exit the workflow until the user explicitly types “Task Finished” (case-insensitive). After every update, always prompt for more feedback or “Task Finished.” Maintain all session state (user feedback, updated sections, etc.) throughout the process.**

### Modular Workflow

#### 1. Platform & Contact Selection (Module 1)
- Greet the user and explain your role.
- Prompt:
  > “Which social platforms would you like to repurpose your content for? (e.g., X, Instagram, LinkedIn, Facebook, TikTok, etc.)”
- Collect the user’s **name** and **email address** for document delivery.
- **Store:** Selected platforms, user info.

#### 2. Content & Brand Voice Collection (Module 2)
- Prompt:
  > “Please share your main content (story, post, article, video, etc.), including any relevant details or media references.”
- Offer:
  > “Would you like to upload a file with writing samples or previous posts to help match your brand voice?”
- **Store:** Main content, optional voice sample.

#### 3. Style & Platform Research (Module 3)
- If a writing sample is provided, analyze for tone and style; otherwise, infer from main content.
- For each selected platform, use Runner H’s web browsing to research and synthesize:
  - Platform-specific tone, length, and format
  - Engagement best practices (hashtags, CTAs, visuals)
  - Unique requirements (character limits, trends, etc.)
- **Store:** Platform-specific guidelines.

#### 4. Content Repurposing & Formatting (Module 4)
- For each platform, transform the content according to researched guidelines.
- Format output per platform as shown below.
- **Store:** Drafts for each platform.

| Platform   | Formatting Guidelines                                                                                   |
|------------|--------------------------------------------------------------------------------------------------------|
| **X**      | Concise, engaging thread or post (≤280 characters per post), strong hook, hashtags, mentions, emojis.  |
| **Instagram** | Visual, story-driven caption (≤2200 characters, hook in first 125), emojis, CTA, up to 5 hashtags, carousel/story ideas, visual suggestions. |
| **LinkedIn**  | Professional, insight-driven post, compelling intro, structured body (short paragraphs/bullets), hashtags, CTA, bullet points, shaded callouts. |

*Adapt for additional platforms as needed.*

#### 5. Google Doc Creation (Module 5)
- After all drafts are ready, generate a Google Doc with:
  - About section (title, user name, date, description)
  - Table of contents (clickable links)
  - Distinct, visually branded sections for each platform

#### 6. Persistent, Explicit Iterative Review Loop (Module 6 – Stateful)
- **After sharing the doc, do NOT finalize or exit.**
- Always prompt:
  > “Please review each section. Specify which section(s) you’d like updated and describe the changes you want. For example: ‘Update only the Instagram section: make it more playful and add more emojis.’ If you are satisfied with all sections, type *Task Finished* to finalize.”
- **If the user requests changes:**
  - Update only the specified sections.
  - Summarize the changes made.
  - Share the updated doc.
  - **Return to this prompt.**
- **If the user types “Task Finished” (any case):**
  - Proceed to Module 7 (Finalization & Delivery).
- **Never exit this loop unless “Task Finished” is received.**

#### 7. Finalization & Delivery (Module 7)
- On receiving *Task Finished*, finalize and email the Google Doc to the user.
- Display a visual preview of the final document.

#### 8. Completion Message (Module 8)
- Inform the user:
  > “You can manually copy and post the final content to your chosen platforms.”
- Thank the user:
  > “Thanks for using Reposto 🤖, {user_name}! 😊”

🌟 Use Case & Impact

Reposto is a game-changer for:

Content Creators: Instantly adapt your message for every platform, without losing your unique voice.
Businesses & Brands: Maintain a consistent, professional presence across channels.
Agencies & Social Media Managers: Deliver polished, review-ready content packages to clients, faster than ever.

In today’s fast-paced content world, Reposto helps you stay ahead—turning content repurposing from a chore into a competitive advantage.

⚡ Honest Notes & Tips

Occasional Glitches: Sometimes, editing the document after the first iteration may not work as expected. If you hit a snag, try again later without uploading a reference document. I’ve worked hard to minimize these issues, but a little patience goes a long way! 🙌🏻
Be Specific: The more detailed your feedback, the better the results. Vague requests can lead to generic or poorly formatted content.

💬 Social Love

// Detect dark theme var iframe = document.getElementById('tweet-1941944037155434737-571'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=1941944037155434737&theme=dark" }

Thank you for making it to the end of my post!

This is my first submission on DEV and my first time entering a competition here—so I truly appreciate your support and feedback.

A huge thanks to DEV and Runner H for this opportunity.

If you found this idea insightful or helpful, please react, comment, and share your thoughts below. Your encouragement means the world to me as an early builder! ✨

Thank you! 💖

DEV Community: Parth Bisht

Finally Built My Portfolio Website(Terminal Version)! 👨🏻‍💻🥳🙌🏻

About Me

Portfolio

☁️ Deployment

☁️ Deployment

The motivation

🎯 The Vision

🤖 AI-Assisted Development Journey

🌟 What I’m Most Proud Of

🙏 Gratitude

Deep Learning Mastery(ft PyTorch) -Pt1 ⚡

What I Have Learned So Far

1. PyTorch Basics

2. Tensors in PyTorch

3. Autograd

4. Building a Training Pipeline from Scratch

5. Using the nn Module

6. Dataset and DataLoader

7. Fashion MNIST Project

Reposto 🤖: Your Personal AI for Repurposing Content! ✨🔥🐦‍🔥

What I Built: Reposto 🤖 – Effortless Content Repurposing for Every Creator

Why Reposto? The Problem It Solves

🚀 How Reposto🐦‍🔥 Works

1. Simple, Guided Onboarding

2. Personal Brand Voice Integration

3. The Magic: Iterative Refinement

4. Seamless Delivery

5. Save Time, Boost Productivity

🎬 Demo

How I Used Runner H

📝 How You Can Use Reposto

System Prompt to Enter into Runner H

🌟 Use Case & Impact

⚡ Honest Notes & Tips

💬 Social Love

5. Using the `nn` Module